aftermarket starters

[RDfella]

Been thinking about the problems some have experienced with aftermarket starters and how they parallel the issues I've worked through on the project. Now the latter runs a Lucas DKX2 distributor (as fitted to Triumph 6T etc). Its baseplate says 14* so it advances 28 crank degrees, but the engine is happiest around 36*, meaning a static advance of 8* is required. Now that would be advanced enough for kick-back with a slow turn-over (as happens at 8* static on my classic Jaguar). But a car's starter has a large spring shockabsorber. There is no room for such or an overload clutch on our aftermarket kits, meaning 8* static advance would risk serious starter damage for us. In my case, I've modified the distributor (easily done) to get around 35 / 36 degrees of distributor advance, meaning only a whisker of static is needed. Downside, that means the engine will be retarded until advance is fully in at around 3,000rpm. Clearly that won't do and so I've fitted lighter distributor weight springs to get it to come in earlier.
At first I worried about the fact an advance curve should match an engine, insomuch as max cyl pressure should occur at just after 20* ATDC. Were I producing 1,000 engines a week or preparing for the IOM TT I would put the engine on a brake and establish the advance needed across the rev range and create an advance mechanism to mirror that. But I'm not and so I didn't. Furthermore, I recalled that on all my bikes with manual advance I only ever retard to start and thereafter they remain on full advance until shut down. I'm sure most Gold Star racers did the same, so I'm not overly concerned about advance curves.
So much for the project, but what about our A series with aftermarket starters? First of all, automatic advance (to guarantee the best spark at starting - something I overlooked recently) is recommended. But we set our timing at full advance with no thought (at least as far as I'm concerned) as to what static advance that leaves us with. Is it enough to risk kick-back? Certainly my GF kicks back - about one kick out of ten. Maybe the advance unit springs are weak - but then fitting stronger ones would delay the advance. Frankly, I think those fitting an 'electric leg' should consider changing to a distributor or electronic ignition, where both full advance and static advance can be set.
Finally, I read that some had experienced broken sprag clutches. On some of these devices it's possible (at overload) for the elements ('rollers') to go over centre, thereby putting immense bursting pressure on the outer race. But there are sprag clutches - one way bearings - available where the elements incorporate a knee which prevents an over-centre situation, thereby reducing the possibility of destruction. Clearly those would be the preferred option (if not already fitted).

[chaterlea25]

Hi RD,
Is it "Great minds think alike"  or "fools seldom differ"
I have had the same thoughts, and have been considering my options
I have in a box a couple of electronic ignition dizzys and their associated gubbins
One is a Boyer which is immediately ruled out as they go full advance on low voltage, which is what it would get when the starter is turning the engine
The second is a "Pazon" (I think?) which has a load of gizmos attached.. I need to research what it actually is??

In the mean time I have a conventional AA unit that I am going to start off with it fitted to the manual advance magneto, so I can retard the timing a touch if I find the static advance is found to be "too far" advanced for the electric starter
I have found that the bike performs better with less advance than the book figure.
This was something one of the members on here found out by putting his various A series bikes on a dyno

At last the Guzzi GTW is off the bench and once a tidy up is done I will be getting to work on the SR electric leg job

John

[Colsbeeza]

That is good info RD and J.O’R,
I am keen to understand the implications of your thoughts, as I have just received my Electric Leg from Stephen McFarlane for my bog standard 1960 Golden Flash, and need to think about these issues.
My main question is how to minimize the risk of kick back, so want to get the timing as accurate as I can WITH WHAT I HAVE, and without access to a dynamometer.
I want to run the bike for some time at BSA’s standard 32Deg or about 5/16” BTDC to settle the motor in before I fit the “Leg”, and want to be happy that kick-backs will be rare before I install it.
I did renovate the carb with new jets and 3½ slide and cleaned out the pilot orifi.
The magneto has been overhauled, and each side is within 10 of the other. I now have a timing light from “The Magneto Guys” and a piston stop and timing disc and have set my timing at 32Deg BTDC, with ATD chocked fully advanced.                                       With a standard ATD range of 11-13Deg, or 24Deg at the crank, the Static timing in theory is therefore about 8Deg.
The old way of setting the timing was to use a cigarette paper or feeler gauge to identify the point at which the points just break.
With this light, you rotate the motor forward until the light just comes on. At this critical point, we then give the ATD bolt a light tap and nip up the bolt. Presumably the points gap is somewhat less than with a cigarette paper or feeler gauge.
So the cig paper or feeler gauge method would give a more advanced setting, likely very different to the Timing Light setting. eg, if the light gives me 32Deg, then a cig paper would come free when points are slightly more open, making the timing say 33-35Deg with Static timing perhaps 9 or 10Deg? This is probably where I have always been setting the timing in the past, and where BSA set by trial and error.? Back in 1960 there were plenty of cigarette papers and very few timing lights. A couple of years ago, it took me ages to find any cig papers.
Once I install the Electric leg, if I were to then set the timing to 30Deg BTDC by the Timing Light to achieve maximum power and torque as per Richard Orabanda’s GF Dynamometer tests  https://www.a7a10.net/forum/index.php?topic=2392.0,
which was a revelation to many of us, then the Static timing should be about 6Deg, which would reduce the risk of kick back and hence damage to the sprag bearings somewhat. But will that be enough?
Does all that sound reasonable??
Col

[groily]

These are interesting Qs.
A typical ATD will start to advance at about 750rpm, if it's got its 2 springs and they're not pre-loaded. And it will hit max advance about 2250 engine rpm. (Obviously, they do vary a bit though  . . .)
At a typical tickover speed of about 800 say, it will be just off the retard stops, with maybe 3° advance at the ATD, 6° crank. At 1000rpm it'll be about 5° ATD and 10° crank off full retard.
If the timing is set set at full advance of 32° (say), and there's 12° on the ATD (say), ie 24° crank, then it's running at about 18° btdc at 1000rpm.

Would it matter if it were running a couple of degrees retarded from those figures at those low speeds? I don't think it would.
So I think, if you wanted to experiment, you could do one of 2 things.

The easy one is get hold of an ATD with more ° on it - say a Triumph version with 15-18° of movement. Set the timing at the same 32° btdc and you get a bit more retard, and I frankly doubt on most engines you'd notice the difference. How many folk actually know whether their ATD is 'BSA spec' or Ariel, Triumph or other spec? Not many of us unless we can test dynamically and know what we're looking for. I have no idea what's on my own A, because I've never bothered to measure it.

The other option would be to dress back the stops for the full retard position on the ATD you've got, to obtain a very few more degrees of retard. As long as the amount removed didn't bugger up the operation of the springs, then, again, I doubt you'd notice the difference. And you'd just maybe reduce the risk of occasional kickback damaging the sprag clutch.

A graph showing what happens with a typical ATD with rising revs is attached, along with specs for the different ATDs used by Lucas in the link. Bear in mind some go the 'wrong way' as per Vincents, but most are for anti-clockwise drive mags.
http://www.brightsparkmagnetos.com/library/LucasATDspecifications.htm

There is no doubt in my mind that Steve M is spot-on in demanding customers don't have manual advance, because the characteristics of a leccy foot are somewhat the opposite of a stout boot, especially when oil is cold and motors turn slowly. A good mag makes almost continuous 5.5mm sparks at about 130rpm fully advanced according to Lucas (or with an ATD of course). They can be got to be a fraction better than that, but that was the model. 5.5mm in fresh air equates to a typical not-daft compression ratio in terms of the energy required to make it happen. 130rpm mag = 260rpm crank  . . . so a leccy foot has some work to do, whatever it's on.

When I think of the hassles people used to have with the original factory-fitted starters on Commandos and so on back in the day (and that with modern coil ignition and all that), I think it's a minor miracle what Steve has achieved, as also for the Vincent people. But never let it be said it's easy, as Keith also said in one of his posts hereabouts.

[Colsbeeza]

Hi Bill,
That is very helpful experience. I hadn't seen those curves before - on Brightsparks under FAQ section. Their description on how they achieved those curves is interesting. One of the curves doesn't make sense though - eg. the Light Blue line. With one bobweight and no springs, why would it remain at fully retarded long after all the others had lifted? That is odd.!
I guess my ATD is their Green Line, as the springs are slightly loose at rest. So your description suits my ATD well. I think you have confirmed my thoughts. However when the time comes, I might add some minor preload to the springs to ensure they stay at zero advance whilst starting - probably fit new ones.
You made a good point about the origins of our ATD's. I remember now I got my ATD from a Swap Meet (autojumble) yonks ago as it had good springs and pivots and 44 teeth. No guarantee it is for the BSA, but almost certainly is.
Re your comment - "There is no doubt in my mind that Steve M is spot-on in demanding customers don't have manual advance, because the characteristics of a leccy foot are somewhat the opposite of a stout boot, especially when oil is cold and motors turn slowly".  I better appreciate the value of a good boot on starting a cold motor. Perhaps on a cold morning, best to kick it over on the centre stand at home if the knee can manage it, then let it warm whilst putting on the gloves and checking the oil return.
Then use the Leccy Boot for the rest of the day. That would give the knee a better time for the rest of the day, and ease the stress of kick starting particularly in Traffic.
My old Hondas (1966 CB72 and 1969 CB450 K2) always preferred a cold kick and then the button when warm, but that was mainly due to wear in the sprag arrangement, massive in size though they were. The sprag wouldn't engage properly unless hot, and then only with no throttle. The rollers always dug into the shaft and created ridges. My Honda experience suggests that continual cold starts with the button may give the BSA Leccy Starter sprag bearing a harder life than necessary?
Thanks Col

[groily]

Yeah, there are some anomalies in how those graphs came out probably Col - we just reported what we recorded as it were - in each state of dilapidation. All we were trying to do there was show that the actual state of springs etc makes less difference than people think sometimes, plus we were suffering from the sort of curiosity that kills cats. But it is true that a decent unit starts moving at about 750 and will be at full advance around 2000rpm.

The irony - perversity almost - in all this is that with a manual mag you get the amount of retard you'd need, with usually about 36° crankshaft available. But fully retarded you get a lousy spark that's far off 'optimal', with it needing maybe 350rpm on the crank to get them flying to Lucas spec. Some older mags had as much as 45° on the crank, so an engine could be persuaded to fire Atdc probably.

I think if your springs are slack at full retard, you're going to be getting some wobble, with the possibility of the timing being a few degrees more advanced than you want as soon as things turn ('cos there's owt to stop it, same reason as the water flows over Niagara Falls). A tiny bit of pre-tension might help if springs can be manipulated to provide it. (And yup, if the right gear's on it, the atd is probably right - but bits do get swapped around, I do it myself when I have to!)

I haven't ever seen the Steve system in operation personally, but I've seen a few Vincents over the years, and have been pretty impressed. However, the last V I saw had an Alton alternator and electronic ignition, coming back to chaterlea's and others' point. Understandable, given the hassles with using (single) mags on V engines anyway, where you find they often fire up on #1, with #2 cutting in when the engine fires.

[Colsbeeza]

Hi Bill,
I didn't know you were behind those tests - I might have been more polite . It is possible that for the Blue line, the dilapidated bobweight got stuck and then suddenly released. Anyway the tests do show that once firing, the advance is not so critical. All the other tests seem good.
The nearest I have got to seeing one of Steve's starters working is on his video at

https://startyourbsa.co.uk/

Seems to turn over quite quickly.!
Cheers
Col

[Greybeard]

The nearest I have got to seeing one of Steve's starters working...
Col

Steve often attends BSAOC weekends. I've seen him use the electric starter on his bike.

[KeithJ]

I had a Vincent for over ten years and fitted an electric leg to it.  One of the best modifictions I made to it.  I was told very strongly not to fit a starter and expect it to solve other problems.  A well set up bike is needed to begin with.  I went through "everything" to make sure the starter would perform correctly and it did.  Timing, carbs, ignition etc.  Another point is Vincents have decompressors so the engine can be spun up and then drop the decompressor.  Puts less load on the starter.  Having tried two mags, on my SR, I have now fitted a Thorspark.  Bike starts second kick for the first start of the day and then first kick.  Unfortunately, I did not refit the starter when I had to strip the gearbox recently so can't say if the Thorspark would have worked better with the electric leg.  My feeling is, it would be better.  The Thorspark uses whatever advance and retard system which is on the bike.  I'm currently using manual.  Another advantage of Thorspark is both cylinders are the same timing as there is no cam ring to worry about.  It also a lot less expensive than having a mag rebuilt.
I believe the manual A/R mag gives a better spark at kickover speeds wheras the Auto A/R is at the wrong part of the magnetic flux in the mag. Excellent point made about checking and adjusting the advance on the A/R units and checking the spring/bobwieghts function correclty.
Keith

[groily]

I believe the manual A/R mag gives a better spark at kickover speeds wheras the Auto A/R is at the wrong part of the magnetic flux in the mag.
Keith

T'other way about Keith.
'Tis why Steve M insists on an ATD on mag-equipped bikes with leccy legs now.

The Vincent thing is a whole different ball game, but post-HRD they had ATDs pretty well on all of them. With decompressors. So some good upsides to compensate for the classic V-engine / single mag weakness.
If I had a V twin , I'd fit electronic ignition too - probably the Thorspark unit as it uses so little current and is miles less pricey than than the B-TH independent spark generator. And probably an Alton alternator too. The Thorspark kit for Vs is pretty versatile - caters for any angle you can think of, being basically a doubled-up version of their single or parallel twin set-up. Still only needs about 1/2A to run, compared to 1/4 for a single (or dead spark twin).

[KeithJ]

Thanks for the correction on the use of manual vs auto A/R on mags.  Will try to remember it.

Also having the auto A/R avoids inadvertantly starting it when not fully retarded and the potential of kickback and damage to sprag clutch.

[Greybeard]

Let's skip starters and go straight to afters 😁

[KeithJ]

Perhaps we will get our just desserts?

[RDfella]

Just a quick update.
Was looking at the fitting instructions for electric starters on Vincents, and they advised no more than 2* of advance be used for starting. Also, that the auto advance unit be checked to ensure the springs are in good condition, thereby ensuring the unit is fully retarded at rest. It is all about preventing kick-back on the starter, as aftermarket units have no capacity to deal with that (apart from breakages).
Essentially, for those fitting electric starters to really any machine this means setting ignition timing to 2* at full retard, and letting the amount of advance as the revs rise be whatever they are. Should the amount of advance be insufficient, then a delve into the auto advance unit is required in order to widen the advance band. I'm running a distributor, but the innards are practically identical to those in an auto unit. I have already widened the advance band (I'd already guessed around 2* would be required rather than the 8 I'd been using) but have not yet verified figures with a strobe as I've been busy strengthening the starter gearbox mounting and replacing the ratchet with a sprag clutch.
Regarding the latter, have to say I was a bit surprised at the torque / tension going through the starter mechanism. Roughly 250 ft/lbs torque at crankshaft and 2 1/2 tons (correction - was working from output shaft revs / dia - actually 2/3 ton) force trying to pull the starter off its mountings. No wonder the mountings were distorting.

[KiwiGF]

Just a quick update.
Was looking at the fitting instructions for electric starters on Vincents, and they advised no more than 2* of advance be used for starting. Also, that the auto advance unit be checked to ensure the springs are in good condition, thereby ensuring the unit is fully retarded at rest. It is all about preventing kick-back on the starter, as aftermarket units have no capacity to deal with that (apart from breakages).
Essentially, for those fitting electric starters to really any machine this means setting ignition timing to 2* at full retard, and letting the amount of advance as the revs rise be whatever they are. Should the amount of advance be insufficient, then a delve into the auto advance unit is required in order to widen the advance band. I'm running a distributor, but the innards are practically identical to those in an auto unit. I have already widened the advance band (I'd already guessed around 2* would be required rather than the 8 I'd been using) but have not yet verified figures with a strobe as I've been busy strengthening the starter gearbox mounting and replacing the ratchet with a sprag clutch.
Regarding the latter, have to say I was a bit surprised at the torque / tension going through the starter mechanism. Roughly 250 ft/lbs torque at crankshaft and 2 1/2 tons force trying to pull the starter off its mountings. No wonder the mountings were distorting.

My recently acquired B44 kick backed like a mule once I had re-commissioned it, positively dangerous I thought, they have a reputation for that, plus being hard to start.

I found weak/loose springs on the advance retard unit, so it was pretty much fully advanced at kickstart speed, also the unit was marked 5 degrees (advance), which equated to 10 degrees retard available at the points, them driving at half speed. Lucas made these AR units with various specs.

The timing is set at 28 btdc on the B44 so I took a dremel to the advance retard unit “slots” to increase it’s range to about 12 degrees, so about 4 btdc at full retard. and also bent the spring tabs to increase spring tension (new springs will be ordered once I find a source).

I checked the timing using a strobe, and it still gets full advance at a fast idle so I reckon even stiffer springs would work ok, and reduce it’s tendency to pink at low revs in high gears.

It no longer kick backs when starting. It does “puff” sometimes after it bounces backwards off a compression stroke, but not badly enough to be called a kick back. I’m not sure if that type of “reversing” action would damage an electric start.

All this got me wondering how the boyer and other electronic ignitions affect the starting of bikes like the B44 and whether some of the reputation for being awkward to start is undeserved. Also if pinking could be caused by an electronic ignition having insufficient advance, or it advances at too low an rpm.

One would like to think the various electronic ignition makers “copy” (or improve) on the original design, but who knows!

I have both Boyer and Pazon ignitions to try out at some point, but I will be watching out for kick backs and use a strobe to get an idea of when full advance kicks in.

The boyer unit is well known for damaging norton commando electric starters, allegedly it “defaults” to fully advanced when the battery voltage drops (as it would during starting.....) and some ignition units allegedly need a couple of revs of the engine to “orientate” themselves before they will provide a spark, which of course would prevent a bike starting first kick!